Audio processing apparatus and control method thereof

1. An audio processing apparatus, comprising: a driving unit; a first microphone whose main acquisition target is sound from outside of the apparatus; a second microphone whose main acquisition target is driving noise from the driving unit, compared with the first microphone; and a noise removing unit configured to generate two-channel audio data in which driving noise made by the driving unit has been reduced, based on a difference between time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone, wherein the noise removing unit includes: a determination unit configured to determine whether or not the driving noise occurred based on the difference between time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone, a correlation processing unit configured to obtain a correlation value between phases of time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone in a case where the determination unit determines that the driving noise occurred, a generation unit configured to generate time-series audio data for which an error of an incident angle of sound from the outside to the first microphone and the second microphone was determined to exceed a preset threshold, out of time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone, based on the correlation value, a first adaptive filter configured to receive inputs of time-series audio data acquired by the first microphone and time-series audio data corresponding to the first microphone and generated by the generation unit, perform adaptive filter processing, and generate audio data of one channel of stereo, and a second adaptive filter configured to receive inputs of time-series audio data acquired by the second microphone and time-series audio data corresponding to the second microphone and generated by the generation unit, perform adaptive filter processing, and generate audio data of the other channel of stereo.

2. The apparatus according to claim 1 , wherein the determination unit determines that time-series audio data acquired by the first microphone includes driving noise in a case where time-series audio data acquired by the second microphone is larger than a preset threshold.

3. The apparatus according to claim 1 , wherein letting that time-series audio data from the first microphone and time-series audio data from the second microphone that have been acquired from the determination unit be M_x[t] and S_x[t], and φ ms = ∑ j = 1 M ⁢ ⁢ M_x ⁡ [ t ] · S_x ⁡ [ t + j ] M ⁢ ⁢ ( j = 0 , 1 , 2 , … ⁢ , M ) , the correlation processing unit determines M at which φms is positive and maximum, as a correlation value.

4. The apparatus according to claim 1 , wherein the first microphone is a microphone whose main target is sound propagated from outside of the apparatus via a first opening portion provided at a predetermined position of a case of the audio processing apparatus, the second microphone is a microphone that converts sound that enters the apparatus via a second opening portion whose area is smaller than that of the first opening portion, into an electrical signal, and a volume of a space between the second microphone and the second opening portion is larger than a volume of space between the first microphone and the first opening portion in order to allow driving noise from the driving unit of the audio processing apparatus to be propagated to the second microphone.

5. The apparatus according to claim 1 , wherein an image capturing unit is provided between the first microphone and the second microphone.

6. A control method of an audio processing apparatus that includes a driving unit, a first microphone whose main acquisition target is sound from outside of the apparatus, and a second microphone whose main acquisition target is driving noise from the driving unit, compared with the first microphone, the method comprising: removing noise so as to generate two-channel audio data in which driving noise made by the driving unit has been reduced, based on a difference between time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone, wherein the removing includes: determining whether or not the driving noise occurred based on the difference between time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone, obtaining a correlation value between phases of time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone in a case where it is determined that the driving noise occurred in this determining, generating time-series audio data for which an error of an incident angle of sound from the outside to the first microphone and the second microphone was determined to exceed a preset threshold, out of time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone, based on the obtained correlation value, receiving inputs of time-series audio data acquired by the first microphone and time-series audio data corresponding to the first microphone and generated in the generating, performing adaptive filter processing, and generating audio data of one channel of stereo, and receiving inputs of time-series audio data acquired by the second microphone and time-series audio data corresponding to the second microphone and generated in the generating, performing adaptive filter processing, and generating audio data of the other channel of stereo.

7. A non-transitory computer-readable storage medium that stores a program of steps of a noise removing method that is read out and executed by a processor in an audio processing apparatus that includes a driving unit, a first microphone whose main acquisition target is sound from outside of the apparatus, and a second microphone whose main acquisition target is driving noise from the driving unit, compared with the first microphone, the method comprising: removing noise so as to generate two-channel audio data in which driving noise made by the driving unit has been reduced, based on a difference between time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone, wherein the removing includes: determining whether or not the driving noise occurred based on the difference between time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone, obtaining a correlation value between phases of time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone in a case where it is determined that the driving noise occurred in this determining, generating time-series audio data for which an error of an incident angle of sound from the outside to the first microphone and the second microphone was determined to exceed a preset threshold, out of time-series audio data acquired by the first microphone and time-series audio data acquired by the second microphone, based on the obtained correlation value, receiving inputs of time-series audio data acquired by the first microphone and time-series audio data corresponding to the first microphone and generated in the generating, performing adaptive filter processing, and generating audio data of one channel of stereo, and receiving inputs of time-series audio data acquired by the second microphone and time-series audio data corresponding to the second microphone and generated in the generating, performing adaptive filter processing, and generating audio data of the other channel of stereo.